Terfenadine, 1-(p-tert-butylphenyl)-4-[4'-(.alpha.-hydroxydiphenylmethyl)-1'-piperidiny l]-butanol is a non-sedating anti-histamine. It is reported to be a specific H.sub.2 -receptor antagonist that is also devoid of any anticholingeric, anti-serotoninergic, and anti-adrenergic effects both in vitro and in vivo. See D. McTavish, K. L. Goa, M. Ferrill, Drugs, 1990, 39, 552; C. R. Kingsolving, N. L. Monroe, A. A. Carr, Pharmacologist, 1973, 15, 221; J. K. Woodward, N. L. Munro, Arzneim-Forsch, 1982, 32, 1154; K. V. Mann, K. J. Tietze, Clin. Pharm. 1989, 6, 331. A great deal of effort has been made investigating structure-activity relationships of terfenadine analogs, and this is reflected in the large number of U.S. patents disclosing this compound and related structures as follows:
U.S. Pat. No. 3,687,956 to Zivkovic PA1 U.S. Pat. No. 3,806,526 to Carr, et. al. PA1 U.S. Pat. No. 3,829,433 to Carr, et. al. PA1 U.S. Pat. No. 3,862,173 to Carr, et. al. PA1 U.S. Pat. No. 3,878,217 to Carr, et. al. PA1 U.S. Pat. No. 3,922,276 to Duncan, et. al. PA1 U.S. Pat. No. 3,931,197 to Carr, et. al. PA1 U.S. Pat. No. 3,941,795 to Carr, et. al. PA1 U.S. Pat. No. 3,946,022 to Carr, et. al. PA1 U.S. Pat. No. 3,956,296 to Duncan, et. al. PA1 U.S. Pat. No. 3,965,257 to Carr, et. al. PA1 U.S. Pat. No. 4,742,175 to Fawcett, et. al. PA1 U.S. Pat. No. 4,254,129 to Carr, et. al. PA1 U.S. Pat. No. 4,254,130 to Carr, et. al. PA1 U.S. Pat. No. 4,285,957 to Carr, et. al. PA1 U.S. Pat. No. 4,285,958 to Carr, et. al. PA1 or R.sub.1 and R.sub.7 taken together form a second bond between the carbon atoms bearing R.sub.1 and R.sub.2 ; PA1 or R.sub.1 and R.sub.2 taken together form a second bond between the carbon atoms bearing R.sub.1 and R.sub.2 ; PA1 R.sub.6 is an alkyl with 1 to 6 carbon atoms with a compound of the formula: ##STR34## under conditions effective to produce a first mixture of regioisomers of the formula: ##STR35## Typically, such acylation is carried out by a Friedel-Crafts reaction, as described above in Process One for Producing Substantially Pure Regioisomers.
Terfenadine has been linked to potentially fatal abnormal heart rhythms in some patients with liver disease or who also take the antifungal drug ketoconazole or the antibiotic erythromycin. In animal and human metabolic studies, terfenadine was shown to undergo high first-pass effect, which results in readily measurable plasma concentrations of the major metabolite 4-[4-[4-(hydroxydiphenylmethyl)-1-piperidinyl]-1-hydroxybutyl]- .alpha.,.alpha.-dimethylphenylacetic acid, also known as terfenadine carboxylic acid metabolite. The terfenadine carboxylic acid metabolite also possesses anti-histamine activity in animal models and may lack the cardiac side effects seen with terfenadine.
Piperidine derivatives related to the terfenadine carboxylic acid metabolite are disclosed in the following U.S. patents:
In these patents, 4-[4-[4(hydroxydiphenylmethyl)-1-piperidinyl]-1-hydroxybutyl]-.alpha.,.alp ha.-dimethylbenzeneacetic acid and related compounds are prepared by alkylation of a substituted piperidine derivative of the formula: ##STR2## with an .omega.-haloalkyl substituted phenyl ketone of the formula: ##STR3## wherein the substituents halo, R.sub.1, R.sub.2, n, z, and R.sub.6 are described in column 6 of U.S. Pat. No. 4,254,130.
It is further described that the .omega.-haloalkyl substituted phenyl ketone wherein Z is hydrogen are prepared by reacting an appropriate straight or branched lower alkyl C.sub.1-6 ester of .alpha.-.alpha.-dimethylphenylacetic acid with the compound of the following formula: ##STR4## under the general conditions of a Friedel-Crafts acylation, wherein halo and m are described in column 11 of U.S. Pat. No. 4,254,129. The reaction is carried out in carbon disulfide as the preferred solvent.
Applicant has discovered that the preparation of ethyl 4-(4-chloro-1-oxobutyl)-.alpha.,.alpha.-dimethylphenylacetate by reaction of 4-chlorobutyryl chloride, aluminum chloride, and ethyl .alpha.,.alpha.-(xocdimethylphenylacetate in carbon disulfide, as described in Example 1 of U.S. Patent Nos. 4,254,130 and 4,285,958 provides an mixture of monosubstituted aromatic regioisomers of the formula: ##STR5## wherein the chlorobutyryl substituent is attached at either of the three aromatic carbons which are meta or para to the dimethylacetate substituent. These regioisomers are not separable by standard techniques of thin layer chromatography, or column chromatography, and low field proton nuclear magnetic resonance spectroscopy is inconclusive in identifying the product of this reaction as a mixture. When the mixture of monosubstituted aromatic regioisomers of the preceding formula is reacted with a piperidine of the formula: ##STR6## a second mixture of aromatic regioisomers is obtained of the formula: ##STR7## wherein the monosubstituted meta, para mixture of regioisomers is obtained.
It is known in the art that a monoalkyl substituent on a benzene ring is ortho, para directing in electropldlic aromatic substitution reactionis such as a Friedel-Crafts reaction. Thus, it would be expected that the Friedel-Crafts reaction of .alpha.-chlorobutyryl chloride with ethyl .alpha.,.alpha.-dimethylphenylacetate would yield predominantly the para substituted product of the formula: ##STR8## because of the electron donating, para-directing character of the dimethylalkyl substituent combined with the steric hindrance associated with reaction of the ortho positions. In practice, the inductive electronic withdrawing effect of the carboxylic ester of ethyl .alpha.,.alpha.-imethylphenylacetate counteracts the expected alkyl electron donating effect, resulting in no significant directing effect for the aromatic substitution reaction. For the described reaction, a statistical mixture of meta to para regioisomers results, with the two meta positions predominating.
The above second mixture of regioisomers can be converted to a third mixture of regioisomers of formula: ##STR9## Although the second mixture of regioisomers and the third mixture of regioisomers can be analyzed by HPLC experiments, a practical separation to obtain gram quantities of substantially pure regioisomers has not been achieved. Each mixture (including the first), would be expected to contain 33% of the para isomer and 67% of the meta isomer. Since these components are inseparable, it has not been possible to obtain either of the regioisomers in each mixture in substantially pure form.